The overall goal of this work is to use computer-based decision logic to assist the practicing physician to appropriately use medical knowledge. We will use patient specific decision analysis to recommend optimal diagnostic and therapeutic procedures for patients with suspected coronary artery disease. Using the HELP system for medical decision making, we have constructed modular decision logic for calculating patient specific probabilities that each of several alternative outcomes will result. Logic also exists to calculate the utility associated with each of these possible outcome states. In addition to developing logic which will enhance the capabilities of the present analysis, we plan to evaluate the system by using it in the offices of six physicians who see patients with coronary artery disease. We also propose to develop and use computer-based logic for diagnosing pulmonary diseases. To provide the detailed descriptions of radiological findings which are needed in the decision logic, radiologists will interact directly with a computer terminal. As a result of this interaction, the radiologist will receive additional clinical information from the comprehensive clinical database which is part of the HELP system. The radiologists feel that this information will allow them to more effectively interpret radiology findings and act as consultants.